This research aims to evaluate different approaches to the organization of care for alcohol disorders, and compare their impact on patterns of cost, utilization and outcomes. We will study how utilization, costs and outcomes of alcohol treatment are affected by two managed care programs: the primary care case manager (PCCM) model, and the capitated specialty vendor model ('carve-out'). The two models differ in the way that they link behavioral health care to medical care, and in the degree of control exercised by the gatekeeper. We expect these design differences to result in different impacts on treatment patterns. The study has three specific aims: 1. Describe and evaluate how service utilization and costs of care for alcohol disorders are affected by managed care, and by the extent to which managed care for alcohol problems is integrated with managed care of other problems. 2. Examine how outcomes for alcohol-disorder patients are affected by managed care, and by the extent to which managed care for alcohol problems is integrated with managed care of other problems. 3. Evaluate whether managed care impacts on cost and utilization differ for women and people of color. To address these aims, we analyze Medicaid data from two states which have implemented the two managed care models in question. Both Massachusetts and Michigan recently made managed care enrollment mandatory for most Medicaid recipients, requiring them to join either PCCMs or capitated plans such as HMOs. However, Massachusetts also 'carved out' the behavioral health services, to be managed by a capitated specialty vendor. By contrast, Michigan did not separate behavioral health from other care, thereby leaving it to be managed by PCCMs. For each of these states, we will collect and analyze claims data for two years before and three years after the start of the managed care intervention. The results will provide insight about the role of organizational arrangements in determining the patterns of treatment for alcohol disorders.